Eating less sodium than usual, most often by cutting salt used in cooking and by eating fewer processed and restaurant foods.
At a glance
Evidence: Randomised trials of sustained sodium reduction lower blood pressure, and the effect is larger in people with higher starting blood pressure.Resting blood pressure in randomised trials · Strong certainty · Randomized intervention
What’s uncertain: Heterogeneity was substantial in the Cochrane meta-analysis (I2=75% for systolic and 68% for diastolic), and several of the trials were small and of short duration.
First step: Review frequent sodium sources such as processed meat, bread, cheese, sauces and takeaway food; this is an editorial starting point.
Before you try it: Ask a clinician before using potassium-based salt substitutes if you have kidney disease or take medicines that raise potassium; avoid self-directed changes with heart failure or diuretic use. All precautions
Read the full takeaway
Blood pressure
The blood-pressure evidence is strong and separates cleanly from everything else. Randomised trials of sustained sodium reduction show significant falls in resting blood pressure in both hypertensive and normotensive adults, with the largest falls in people with higher starting blood pressure, and one dose-response meta-analysis found an approximately linear relationship across the whole range of intake studied rather than a threshold.
Cardiovascular events and mortality
How far that translates into fewer heart attacks, strokes or deaths is much less settled. The randomised trial evidence on cardiovascular events and mortality is small, imprecise and driven by a single institutional trial, and cohort studies disagree with each other: some report higher cardiovascular risk at high sodium intakes, including a large cohort in low-income US adults, while others report a J-shaped or inverse pattern.
Measurement
Measurement is a real part of that disagreement. Validation studies in 970 stroke patients and 807 older hypertensive adults show that spot-urine formulas can be nearly unbiased on average in a specific population while still agreeing poorly with measured 24-hour sodium for any one person, which is exactly the kind of error that could contribute to an apparent J-shape without demonstrating that one exists.
Overall conclusion
This record therefore supports reducing high sodium intake, especially if your blood pressure is raised, and does not support a universal sodium or salt target, nor a confident claim that cutting sodium will extend your life.
Outcome assessments based on research, including its uncertainties and limitations.
Resting blood pressure in randomised trials
Finding: Benefit supportedCertainty: Strongstudies found a benefit from doing this
Evidence basis: randomised trials
Randomised trials of sustained sodium reduction lower blood pressure, and the effect is larger in people with higher starting blood pressure.
Population: Adults in randomised trials: 3,230 participants in 34 trials of at least 4 weeks in the Cochrane review of longer-term modest salt reduction; 85 trials in the dose-response meta-analysis, conducted in participants with hypertension (n=65 trials), without hypertension (n=11) or a combination (n=9); 195 references, with blood-pressure outcomes stratified by race, in the broader Cochrane review of low versus high sodium diets.
Research details
Summary
Randomised trials of sustained sodium reduction lower blood pressure, and the effect is larger in people with higher starting blood pressure. A Cochrane review of 34 trials with 3,230 participants and at least four weeks of intervention found a mean urinary sodium change of -75 mmol/24 h, equivalent to a reduction of 4.4 g/day of salt, with a mean fall of 4.18 mmHg systolic (95% CI -5.18 to -3.18) and 2.06 mmHg diastolic (95% CI -2.67 to -1.45); the fall was 5.39/2.82 mmHg in hypertensive participants and 2.42/1.00 mmHg in normotensive participants. A separate dose-response meta-analysis of 85 trials (sodium intake 0.4 to 7.6 g/day, follow-up 4 weeks to 36 months) found an approximately linear relationship between achieved sodium intake and both systolic and diastolic blood pressure, with no flattening at either end of the range and a steeper fall in participants with hypertension. A third Cochrane review covering 195 references and 89 to 96 blood-pressure comparisons reported smaller average effects alongside consistent hormonal changes.
Why this certainty
Rated Strong for the direction of the blood-pressure effect: multiple independent randomised syntheses agree, the effect is larger where baseline blood pressure is higher, a dose-response relationship is present, and the Cochrane review graded most of the blood-pressure and hormone evidence as high quality. It is not rated higher, and it is deliberately not extended to clinical outcomes, because the outcome is a surrogate, heterogeneity is substantial, and the trials cannot show whether the blood-pressure fall translates into fewer events.
Limitations
Heterogeneity was substantial in the Cochrane meta-analysis (I2=75% for systolic and 68% for diastolic), and several of the trials were small and of short duration. Detection bias was the largest bias contrast across studies, although trials at low risk of detection bias did not differ from the rest. Effects on blood pressure are a surrogate outcome; they are not evidence that clinical events are reduced, and the trials were not designed or powered to show that. The dose-response analysis covers sodium intakes as low as 0.4 g/day, far below any usual diet, so it does not establish the safety or benefit of extreme restriction.
Effect estimate
Cochrane 34 trials (urinary sodium -75 mmol/24 h, equivalent to -4.4 g/day salt): systolic blood pressure -4.18 mmHg (95% CI -5.18 to -3.18, I2=75%) and diastolic -2.06 mmHg (95% CI -2.67 to -1.45, I2=68%); hypertensive subgroup -5.39/-2.82 mmHg and normotensive subgroup -2.42/-1.00 mmHg. he-2013 meta-regression: a 100 mmol/24 h reduction in urinary sodium (about 6 g/day salt) was associated with a 5.8 mmHg fall in systolic blood pressure (95% CI 2.5 to 9.2, P=0.001).
Cardiovascular events and mortality in randomised trials
Finding: Mixed evidenceCertainty: Limitedstudies disagree, or the effect depends on who is studied
Evidence basis: randomised trials
The randomised evidence on hard clinical outcomes is small and inconsistent, and it does not carry the confidence of the blood-pressure evidence.
Population: Adults in randomised trials with at least six months of follow-up: 3,518 normotensive participants and 3,766 hypertensive or mixed participants across eight trials in the Cochrane review, with end-of-trial follow-up of 6 to 36 months and post-trial observational follow-up up to 12.7 years. The TOHP follow-up covered adults aged 30 to 54 years with prehypertension across 10 and nine clinic sites in the United States.
Research details
Summary
The randomised evidence on hard clinical outcomes is small and inconsistent, and it does not carry the confidence of the blood-pressure evidence. One long-term observational follow-up of two randomised sodium-reduction trials in adults with prehypertension (744 participants in TOHP I and 2,382 in TOHP II, followed 10 to 15 years) found a 25% lower risk of a cardiovascular event in the intervention groups (relative risk 0.75, 95% CI 0.57 to 0.99) and 30% lower after further adjustment (0.70, 95% CI 0.53 to 0.94), with 67 deaths showing no significant difference (0.80, 95% CI 0.51 to 1.26). A Cochrane review of 8 randomised trials with at least 6 months of follow-up found imprecise all-cause mortality estimates with no evidence of benefit in either normotensive or hypertensive participants, weak evidence for cardiovascular mortality in hypertensive participants, and weak evidence for fewer cardiovascular events that was driven by one trial among retirement home residents in which salt was reduced in the kitchens rather than by individual behaviour change. An earlier version of that Cochrane review also found that salt restriction increased all-cause death in the 232 participants with congestive heart failure (relative risk 2.59, 95% CI 1.04 to 6.44, 21 deaths) and called for further trial evidence to confirm whether restriction is harmful in that group; the heart failure trial was not carried into the 2014 update, so this signal is unresolved rather than replicated or refuted. The review authors concluded that there is insufficient power to confirm clinically important effects and that their estimates of benefit were larger than would be predicted from the small blood pressure reductions achieved. A separate umbrella review of 56 systematic reviews reported moderate to high certainty that lower sodium intake reduces systolic blood pressure by between 2.00 and 8.69 mmHg, and that salt substitutes conferred a small but important absolute reduction in all-cause mortality (12 fewer per 1,000) and cardiovascular mortality (9 fewer per 1,000) with little to no effect on stroke (1 fewer per 1,000); the abstract retrieved does not state the constituent trials, populations or sodium reduction achieved, so those figures cannot be checked further from the record.
Why this certainty
Rated Limited: the trials with hard outcomes are few and underpowered, all-cause mortality showed no clear effect in either population, the only statistically robust pooled result depended on one trial delivered at institutional level, and the authors of the Cochrane review explicitly concluded that the evidence is insufficient to confirm clinically important effects. It is rated above Very limited because two independent lines of trial-based evidence point in the same direction (the Cochrane review and the TOHP follow-up), but it is deliberately not rated Moderate because the data cannot support a confident claim about mortality.
Limitations
Event numbers are low: 60 to 79 deaths in normotensive participants and 565 to 674 in hypertensive participants across the whole Cochrane review, so confidence intervals are wide enough to include both meaningful benefit and meaningful harm. The TOHP analysis used an observational follow-up of trial cohorts rather than randomised allocation over the full period, and its authors describe the conclusion cautiously as 'may also reduce' cardiovascular risk; the abstract does not report whether adverse events or baseline differences between groups were compared, so that remains unchecked from the material retrieved. The one pooled cardiovascular-event result that excluded the null (RR 0.77, 95% CI 0.63 to 0.95, six trials, n=5,912) was driven by a single institutional trial, and the 2014 Cochrane review also stated that its estimates of clinical benefit were larger than the achieved blood pressure reductions would predict, which is a consistency problem rather than reassurance. The heart failure harm signal comes from the 2011 version of the review (about 232 participants in one trial) and is not reported in the 2014 update, so it should be read as an unresolved safety flag rather than a confirmed harm of sodium reduction.
Effect estimate
TOHP I and II follow-up: cardiovascular events relative risk 0.75 (95% CI 0.57 to 0.99, P=0.04), 0.70 (95% CI 0.53 to 0.94) after further adjustment, and all-cause death 0.80 (95% CI 0.51 to 1.26, P=0.34) with 67 deaths. Cochrane review: all-cause mortality in normotensive participants at end of trial RR 0.67 (95% CI 0.40 to 1.12) and at longest follow-up RR 0.90 (95% CI 0.58 to 1.40); in hypertensive participants RR 1.00 (95% CI 0.86 to 1.15) and RR 0.99 (95% CI 0.87 to 1.14); pooled cardiovascular events across six trials RR 0.77 (95% CI 0.63 to 0.95, n=5,912), driven by one trial in retirement home residents; cardiovascular mortality in hypertensive participants RR 0.67 (95% CI 0.45 to 1.01). Earlier Cochrane version (2011): all-cause death with salt restriction in congestive heart failure RR 2.59 (95% CI 1.04 to 6.44, 21 deaths).
Long-term cardiovascular events and mortality in cohort studies, and the measurement problem behind the disagreement
Finding: Mixed evidenceCertainty: Limitedstudies disagree, or the effect depends on who is studied
Evidence basis: observational studies (association, not proof of cause)
Cohort findings on sodium and hard outcomes genuinely conflict, and how sodium is measured is part of the conflict rather than a technicality.
Population: Prospective cohort studies and pooled cohorts worldwide: 20 prospective studies in the dose-response meta-analysis; 133,118 individuals (63,559 with hypertension and 69,559 without, median age 55 years) from 49 countries in four prospective studies pooled in the PURE analysis, median follow-up 4.2 years; 64,329 Southern Community Cohort Study participants aged 40 to 79 years recruited at community health centres in 12 southeastern US states, 71.8% Black and 60.9% female, median follow-up 13.8 years. Measurement validation studies now cover 970 stroke patients in Shanxi province, China (689 development and 281 validation, mean age 64.5 years); 807 older hypertensive adults in rural Ningxia, China; 99 healthy Chinese volunteers aged 21 to 57 years; 23 community-dwelling South African adults; and 12 healthy Dutch volunteers studied on three controlled sodium diets.
Research details
Summary
Cohort findings on sodium and hard outcomes genuinely conflict, and how sodium is measured is part of the conflict rather than a technicality. A dose-response meta-analysis of 20 prospective studies found a significant non-linear association between urinary sodium excretion and cardiovascular mortality with a relatively steep slope above 2,400 mg/day, but found no significant association in studies that used dietary sodium intake, and no significant non-linear association with all-cause mortality; a 100 mg/day increment gave an effect size of 1.01 (95% CI 0.97 to 1.05) for cardiovascular mortality and 1.01 (95% CI 1.00 to 1.02) for all-cause mortality. In a pooled analysis of 133,118 people from 49 countries, sodium excretion of 7 g/day or more was associated with higher risk in participants with hypertension (HR 1.23, 95% CI 1.11 to 1.37) and excretion below 3 g/day was associated with higher risk in both those with hypertension (HR 1.34, 95% CI 1.23 to 1.47) and those without (HR 1.26, 95% CI 1.10 to 1.45), while higher excretion was not associated with risk in participants without hypertension (HR 0.90, 95% CI 0.76 to 1.08). In a cohort of 64,329 predominantly low-income Black and White American adults followed a median of 13.8 years, higher dietary sodium was associated with cardiovascular and coronary heart disease death, and excess sodium was estimated to account for 10% of cardiovascular, 13% of coronary heart disease and 30% of heart failure deaths. Measurement is where the disagreement partly lives, and two large validation studies now show it is not a technicality: spot-urine estimation methods are being tested against measured 24-hour collections in stroke patients and in older hypertensive adults, and even the best-performing formulas in those studies were accurate for a minority of individuals.
Why this certainty
Rated Limited and recorded as Mixed/uncertain because the cohort literature does not agree on direction at either end of the intake range, the pooled estimates are highly heterogeneous, and the exposure measurement that drives the disagreement is demonstrably inaccurate at the individual level in validation studies. Positive findings from large dietary-assessment cohorts and the non-linear urinary-sodium signal sit alongside a null association in the same meta-analysis when dietary intake is used, and alongside an inverse or J-shaped pattern in other pooled analyses. The evidence is not strong enough for a confident claim in either direction, and the blood-pressure certainty was deliberately not carried over to these hard outcomes.
Limitations
All of these designs are observational at the level of the exposure-outcome question, and each has a different exposure measurement problem. The PURE analysis estimated 24-hour urinary sodium from a single spot or morning urine as a group-level measure, and a published erratum exists. The Southern Community Cohort Study measured intake by food frequency questionnaire rather than by urinary biomarker, so it shares the misreporting problems of dietary assessment. The dose-response meta-analysis found the association only in urinary-based studies, and subgroup analysis located it only in studies conducted in the USA. Validation studies that used reference 24-hour collections showed substantial bias and poor precision: over 40% of Dutch participants differed from measured 24-hour sodium by more than 1,000 mg/day; all three prediction equations grossly overestimated excretion in South African adults and detected almost nobody with intake at or below 5 g/day of salt; individual misclassification exceeded 60% in Chinese volunteers; and overestimation at low intakes with underestimation at high intakes was a recurring pattern. The two larger studies do not overturn that picture so much as show how narrow the exceptions are. In 807 older hypertensive adults in rural Ningxia, a locally developed equation and the INTERSALT method showed no statistically significant mean bias against measured 24-hour excretion while the Kawasaki and Tanaka methods were grossly biased: the paper prints those mean differences as 40.81 and 18.28 g/d against its own measured mean excretion of about 3 g/d, which is evidently a unit error in the source, so this entry relies only on the direction and the gross size of that discrepancy and not on the printed unit. The intraclass correlation for even the best method was 0.31, which is poor agreement at the individual level, and the authors restrict their conclusion to this population and warn that extrapolation needs care. In 970 stroke patients in Shanxi, a newly developed equation achieved a mean bias of 5.17 mmol/day and the best P30 in the comparison, and the authors still describe that accuracy as low, with only 59.43% of estimates falling within 30% of the measured value. These studies test the validity of spot-urine estimation; they do not measure the shape of the sodium-mortality relationship, and they cannot be used to argue that any observed J-shape is real or that it is an artefact. What they support is the narrower point that misclassification of this size could contribute to an apparent J-shaped or inverse association even if the true relationship were monotonic. Pooled estimates also carry high heterogeneity, with I2 of 98.4% for cardiovascular mortality and 89.2% for all-cause mortality in the meta-analysis of 100 mg/day increments.
Effect estimate
Per 100 mg/day increment in sodium: cardiovascular mortality 1.01 (95% CI 0.97 to 1.05, I2=98.4%) and all-cause mortality 1.01 (95% CI 1.00 to 1.02, I2=89.2%); significant non-linear association with cardiovascular mortality only in studies assessing urinary sodium excretion, with a relatively steep slope above 2,400 mg/day, and no significant association in studies using dietary sodium intake (P=0.61). PURE pooled analysis versus 4 to 5 g/day sodium: 7 g/day or more in hypertension HR 1.23 (95% CI 1.11 to 1.37); below 3 g/day HR 1.34 (95% CI 1.23 to 1.47) in hypertension and HR 1.26 (95% CI 1.10 to 1.45) without hypertension; 7 g/day or more without hypertension HR 0.90 (95% CI 0.76 to 1.08). Southern Community Cohort Study per 1,000 mg/day: cardiovascular death HR 1.07 (95% CI 1.03 to 1.10) in Black participants and 1.08 (95% CI 1.02 to 1.14) in White participants; coronary heart disease death HR 1.08 (95% CI 1.02 to 1.14) and 1.13 (95% CI 1.03 to 1.23) respectively; population-attributable risk 10% of cardiovascular, 13% of coronary heart disease and 30% of heart failure deaths. Validation studies (measurement validity only). 807 older hypertensive adults: printed mean bias 0.09 and -0.07 against measured 24-hour excretion, both non-significant, with intraclass correlation 0.31 (95% CI 0.22 to 0.39) for the best method. Those two bias figures are reproduced exactly as the paper prints them and nothing more is claimed for them: the paper labels those rows g/day while reporting a measured mean of about 3 g/day, so the unit is evidently wrong, and its printed intervals do not contain their own point estimates. The measurement point this entry relies on is the poor individual-level agreement, not those means. 970 stroke patients: new equation mean bias 5.17 mmol/day (95% CI -1.93 to 12.27) with P30 59.43% against 49.47% for Kawasaki, 50.53% for INTERSALT and 56.58% for Tanaka.
No sourced public statements added for this habit yet. This does not mean no one discusses it.
Practical steps and precautions
How to try it
Start with the biggest sources if you want one change: processed meats, bread, cheese, sauces, stock cubes and restaurant or takeaway food usually contribute more sodium than the salt shaker at home (editorial starting point, not a measured dose).
Read labels and compare products by sodium per 100 g, then keep the lower-sodium version of whatever you buy regularly rather than changing your whole diet at once.
Cook more from basic ingredients and season with herbs, spices, garlic, chilli and citrus instead of relying on salt; the blood-pressure trials delivered a sustained reduction in intake rather than a temporary one.
If you use table salt, WHO conditionally suggests replacing regular salt with a potassium-containing lower-sodium substitute for eligible adults. This is a substitution, not permission to add more salt; it complements reducing sodium in packaged foods. Check the exclusions and medicines first.
Precautions
There is no settled universal sodium or salt target. In one pooled cohort analysis both the highest and the lowest sodium excretion groups had higher cardiovascular risk than moderate intake, so the goal is reducing a high intake rather than getting as low as possible. The WHO salt-substitute recommendation excludes children, pregnant people, and people with kidney impairment or other conditions that compromise potassium excretion. Ask a clinician before use if a medicine can raise potassium.
Salt substitutes sold as low-sodium products usually replace some sodium with potassium. They can raise blood potassium to dangerous levels in kidney disease and with medicines such as ACE inhibitors, angiotensin receptor blockers, potassium-sparing diuretics or aldosterone antagonists; ask a clinician before using one in those situations.
Do not change your sodium intake on your own if you have heart failure or are taking diuretics. The 2011 Cochrane review found that salt restriction increased all-cause death in the one small trial in people with congestive heart failure (232 participants, 21 deaths, RR 2.59, 95% CI 1.04 to 6.44); that trial was not carried into the 2014 update, so the signal is unresolved rather than confirmed.
Sustained sodium reduction causes small physiological increases in plasma renin activity, aldosterone and noradrenaline, which is expected rather than a symptom, but it is the main proposed mechanism by which benefits on blood pressure might be partly offset.
References
Effect of longer-term modest salt reduction on blood pressure
He FJ, Li J, MacGregor GA · Cochrane Database of Systematic Reviews (also published as BMJ 2013;346:f1325) · 2013
Cochrane systematic review and meta-analysis of randomised trials with a modest reduction in salt intake and duration of at least four weeks, with random-effects meta-analysis, subgroup analysis and meta-regressionVerified from abstract
Blood Pressure Effects of Sodium Reduction: Dose-Response Meta-Analysis of Experimental Studies
Filippini T, Malavolti M, Whelton PK, Naska A, Orsini N, Vinceti M · Circulation · 2021
Dose-response meta-analysis of experimental studies using a one-stage cubic spline mixed-effects model, restricted to trials with at least 4 weeks of follow-up and 24-hour urinary sodium measurementsVerified from abstract
Effects of low sodium diet versus high sodium diet on blood pressure, renin, aldosterone, catecholamines, cholesterol, and triglyceride
Graudal NA, Hubeck-Graudal T, Jurgens G · Cochrane Database of Systematic Reviews · 2020
Cochrane systematic review and meta-analysis of randomised trials comparing low-sodium with high-sodium diets, with GRADE certainty assessment and blood-pressure outcomes stratified by raceVerified from abstract
Long term effects of dietary sodium reduction on cardiovascular disease outcomes: observational follow-up of the trials of hypertension prevention (TOHP)
Cook NR, Cutler JA, Obarzanek E, Buring JE, Rexrode KM, Kumanyika SK, Appel LJ, Whelton PK · BMJ · 2007
Observational long-term follow-up of two completed randomised trials of dietary sodium reduction, assessed 10 to 15 years after the original trialsVerified from abstract
Reduced dietary salt for the prevention of cardiovascular disease
Adler AJ, Taylor F, Martin N, Gottlieb S, Taylor RS, Ebrahim S · Cochrane Database of Systematic Reviews · 2014
Cochrane systematic review and meta-analysis of randomised trials with at least six months of follow-up in which the intervention was reduced dietary salt through advice or low-sodium salt substitution, with mortality or cardiovascular morbidity outcomesVerified from abstract
Impact of sodium intake on blood pressure, mortality and major cardiovascular events: an umbrella review of systematic reviews and meta-analyses
Yang Q, Vernooij RWM, Zhu H, Nesrallah G, Bai C, Wang Q, Li Y, Xia D, Bala MM, Warzecha S, Sun M, Jayedi A, Shab-Bidar S, Pan B, Tian J, Yang K, Ge L, Johnston BC · Critical Reviews in Food Science and Nutrition · 2025
Umbrella review of systematic reviews with meta-analyses, appraised with a modified and more stringent AMSTAR 2 instrument and GRADE, with absolute risk reductions computed across three cardiovascular risk strataVerified from abstract
Reduced dietary salt for the prevention of cardiovascular disease (2011 version)
Taylor RS, Ashton KE, Moxham T, Hooper L, Ebrahim S · Cochrane Database of Systematic Reviews · 2011
Cochrane systematic review and meta-analysis of randomised trials with at least six months of follow-up of reduced dietary salt intake, with mortality or cardiovascular morbidity outcomes (earlier version, later updated and withdrawn)Verified from abstract
Dose-response association of dietary sodium intake with all-cause and cardiovascular mortality: a systematic review and meta-analysis of prospective studies
Milajerdi A, Djafarian K, Shab-Bidar S · Public Health Nutrition · 2019
Systematic review and dose-response meta-analysis of prospective cohort studies, examining linear and non-linear associations between sodium intake and all-cause and cardiovascular mortalityVerified from abstract
Associations of urinary sodium excretion with cardiovascular events in individuals with and without hypertension: a pooled analysis of data from four studies
Mente A, O'Donnell M, Rangarajan S, Dagenais G, Lear S, McQueen M, Diaz R, Avezum A, Lopez-Jaramillo P, Lanas F, Li W, Lu Y, Yi S, Rensheng L, Iqbal R, Mony P, Yusuf R, Yusoff K, Szuba A, Oguz A, Rosengren A, Bahonar A, Yusufali A, Schutte AE, Chifamba J, Mann JF, Anand SS, Teo K, Yusuf S, PURE, EPIDREAM and ONTARGET/TRANSCEND Investigators · Lancet · 2016
Pooled analysis of four large prospective studies (PURE, EPIDREAM, ONTARGET and TRANSCEND) relating estimated 24-hour urinary sodium excretion, used as a group-level measure of intake, to a composite of death and major cardiovascular events and to blood pressure, stratified by hypertension statusVerified from abstract
Sodium Intake and Cause-Specific Mortality Among Predominantly Low-Income Black and White US Residents
Yoon HS, Cai Q, Yang JJ, Lipworth L, Cai H, Yu D, Steinwandel MD, Gupta DK, Blot WJ, Zheng W, Shu XO · JAMA Network Open · 2024
Prospective cohort study with multivariable-adjusted Cox regression, assessment of non-linear associations and estimation of population-attributable risk for mortality associated with excess sodium intakeVerified from abstract
Validation of spot urine in estimating 24-h urinary sodium, potassium and sodium-to-potassium ratio during three different sodium diets in healthy adults
Validation study in which 12 healthy volunteers followed three dietary sodium targets for three consecutive weeks each and measured daily spot urine, with 24-hour urine collection on day 7 of each week as the reference standardVerified from abstract
Correcting for Intra-Individual Variability in Sodium Excretion in Spot Urine Samples Does Not Improve the Ability to Predict 24 h Urinary Sodium Excretion
Charlton KE, Schutte AE, Wepener L, Corso B, Kowal P, Ware LJ · Nutrients · 2020
Validation study in 23 community-dwelling adults who collected 24-hour and spot early morning urine samples over three consecutive days, comparing INTERSALT, Tanaka and Kawasaki prediction equations with and without adjustment for intra-individual variability against measured 24-hour urinary sodiumVerified from abstract
Assessing whether a spot urine specimen can predict 24-h urinary sodium excretion accurately: a validation study
Zhang Y, Peng Y, Li K, Peng X · Journal of Hypertension · 2019
Validation study in 99 healthy volunteers aged 21 to 57 years in which all spot urines within 24 hours were collected alongside measured 24-hour urinary sodium excretion, and five published equations were compared against the measured valueVerified from abstract
Estimating 24-h urinary sodium excretion from casual spot urine specimen among hypertensive patients in Northwest China: the Salt Substitute and Stroke Study.
Zhao Y, Liu W, Liu S, Li X, Yin T, Liu X, Wang F, Chang X, Zhang T, Tian M, Zhang Y · Public health nutrition · 2021
Validation study nested in the Salt Substitute and Stroke Study, a cluster-randomised trial. Older hypertensive participants in rural Ningxia provided paired casual spot and 24-hour urine samples between January 2015 and February 2017; a sex-specific equation was developed by linear forward stepwise regression and about half the participants were randomly selected to validate it, against the INTERSALT, Kawasaki and Tanaka methods. Both the abstract and the open-access full text were retrieved during this pass (PubMed 32345383, PMC11574822); `accessLevel` records that the verifiedFacts below are quoted from the abstract, and the full-text readings the entry relies on are recorded here so the provenance is explicit rather than invisible. Full text, study flow: 1,120 invited, 162 (14.46%) refused, 64 (6.7%) could not provide a casual spot sample, 73 (7.6%) provided an unqualified 24-hour sample and 14 (1.5%) were excluded for missing data, leaving 807 (84.2%) in the analysis. Full text, Results and Tables 2 and 3: the mean difference against the laboratory method was -0.07 g/d (95% CI -0.10 to 0.25) for INTERSALT and 0.09 g/d (95% CI -0.22 to 0.04) for the new equation, while Kawasaki and Tanaka overestimated by 40.81 g/d (95% CI 39.27 to 42.35) and 18.28 g/d (95% CI 17.73 to 18.83); the intraclass correlation coefficient was 0.31 (95% CI 0.22 to 0.39) for the new equation, 0.10 (95% CI 0.00 to 0.24) for INTERSALT, 0.10 (95% CI 0.00 to 0.25) for Kawasaki and 0.27 (95% CI 0.12 to 0.39) for Tanaka. Unit caveat recorded rather than smoothed over: those mean differences are printed as g/d against a measured mean 24-hour sodium excretion of 3.03 g/d (sd 1.52) in the same table, which would mean a fourfold to fourteenfold overestimate of a quantity the table itself reports as about 3 g/d; a salt intake of 43.85 g/day is implausible, the manuscript elsewhere reports sodium concentrations in mmol/L and states the molar mass of sodium as 23 mg/mmol, and the intended unit in those rows is evidently milligrams per day. The values are reproduced as printed because that is the retrieved text, and the entry cites only the direction and magnitude of the discrepancy, not a unit-correct figure. Print issue 2021 (volume 24, issue 4); the publisher's online-first date is 2020, which is the year the study is often cited as.Verified from abstract
A Method for Estimating 24 h Urinary Sodium and Potassium Excretion by Spot Urine Specimen in Stroke Patients.
Wu B, Yang H, Ren X, Qi Z, Tang S, Yin X, Huang L, Tian M, Wu Y, Feng X, Li Z · Nutrients · 2022
Validation study in stroke patients: paired spot and 24-hour urine samples were used to develop sex-specific equations for 24-hour urinary sodium and potassium excretion by stepwise multivariate linear regression in a training set, then internally validated against published formulas at the population and individual levels. Bibliographic identity was confirmed from Europe PMC and the publisher during this pass rather than from the summary that named the study: PMID 36235755, PMCID PMC9573759, DOI 10.3390/nu14194105, Nutrients 2022;14(19):4105. Both the abstract and the open-access full text were retrieved; `accessLevel` records that the verifiedFacts below are quoted from the abstract, and the full-text readings the entry relies on are recorded here so the provenance is explicit. Full text, Results: 1,287 participants were initially included and 317 were excluded, leaving 970 (75.4%) stroke patients, 689 in the training cohort and 281 in the testing cohort; mean age 64.53 +/- 7.87 years and 531 (54.74%) male. Full text, Table 4: for 24-hour urinary sodium the new formula's mean bias was 5.17 mmol/day (95% CI -1.93 to 12.27) with a Pearson correlation of 0.45 and an intraclass correlation coefficient of 0.35 (95% CI 0.24 to 0.45), and its P30 - the proportion of estimates within 30% of the measured value - was 59.43% against 49.47% for Kawasaki, 50.53% for INTERSALT and 56.58% for Tanaka. Full text, Discussion: the P30 indicates that the accuracy of the new method is low, and the authors present the formula as a meaningful exploration rather than an established method.Verified from abstract
Revision 7 · Approved for publication · Codex (publication approval explicitly authorized by site owner; category migration)
Prepared and checked with AI. Publication approval does not imply independent expert review. Request a correction.
Applied the audit corrections for sodium-reduction. (1) Deleted two uncited validation studies from the population text ('970 stroke patients in China' and '807 older hypertensive participants in rural China'). Neither existed in this entry's source list or anywhere in content/sources.json, and repeated Europe PMC searches for spot-urine validation studies in stroke or rural Chinese hypertensive populations returned no match, so the sentence now describes only the three validation sources actually cited (12, 23 and 99 participants). UNRESOLVED: if a human can identify those two papers they should be added and the broader claim restored, because the measurement argument is currently supported by much smaller samples than the text implied. (2) Made the attribution of the 100 mmol/24 h meta-regression figure explicit as he-2013; it previously sat inside the blood-pressure effect estimate directly after Filippini figures and would have been read as Filippini's. (3) Rewrote precaution 3 so it says the heart-failure harm signal comes from the 2011 version of the Cochrane review and was not carried into the 2014 update; the card-facing precaution previously presented it as an unresolved signal without saying it came from the superseded review. (4) Replaced the card summary, which stated only that the blood-pressure evidence is strong, with one that states the direction, the certainty and the clinical-outcome caveat. (5) Softened implementation step 4: the claim that salt substitutes have 'direct trial support for clinical outcomes' overstated what yang-2025 shows, since it is an umbrella review of reviews whose abstract does not state the constituent trials, populations or sodium reduction achieved. (6) The audit noted that yang-2025 was listed for the cardiovascular-events outcome but contributed no number; it is now used, with its umbrella-review blood-pressure range and salt-substitute absolute risk reductions added to that outcome's summary.
Second reviewer must re-check: (a) whether the two missing validation studies can be identified and cited; (b) whether the salt-substitute mortality benefit applies to people without hypertension, which the retrieved umbrella review abstract cannot resolve; (c) that the Strong blood-pressure certainty is still not carried into any clinical-outcome outcome. The two validation studies deleted by an earlier pass were identified, retrieved and added: Zhao Y et al, Public Health Nutrition 2021;24(4), DOI 10.1017/S1368980019005019 (807 of 1,120 invited older hypertensive adults in rural Ningxia, the Salt Substitute and Stroke Study; PMID 32345383, PMCID PMC11574822) and Wu B et al, Nutrients 2022;14(19):4105, DOI 10.3390/nu14194105 (970 stroke patients in Shanxi, equations developed in 689 and internally validated in 281; PMID 36235755, PMCID PMC9573759). Both were read as full text as well as abstract, and both are cited from the cohort-mortality-and-measurement outcome, which is the one that carries the measurement-error discussion; the earlier note's UNRESOLVED item is closed. What they add is narrow and is stated as such: they test whether spot-urine formulas estimate 24-hour sodium accurately, and they do not measure the shape of the sodium-mortality relationship. The most useful thing they establish is that near-zero mean bias and poor individual agreement coexist - 0.09 g/d and -0.07 g/d mean bias with an intraclass correlation of 0.31 in the Ningxia cohort, and a mean bias of 5.17 mmol/day with only 59.43% of estimates within 30% of measured in the stroke cohort, which the authors of that study themselves call low accuracy. The wording that measurement error 'can create' an apparent J-shape was softened to 'could contribute to', because no retrieved source demonstrates the distortion. The population and limitations text no longer implies that the measurement argument rests on 12-, 23- and 99-participant studies, and it now says explicitly that these two studies cannot be used to argue the J-shape is real or that it is an artefact.
A second reviewer must check: (a) whether the intraclass correlation in the Ningxia study is the right statistic to quote for individual-level agreement, or whether the Bland-Altman limits of agreement would be better; (b) whether citing validation studies from a stroke population and an older hypertensive population in a general-population outcome overstates how far the measurement problem generalises; (c) that the abstract-versus-full-text provenance on both records is accurate, since the abstracts are quotable and the full texts were also read for the flow and agreement figures.
Precision correction to the measurement-validation discussion, made after re-reading both validation studies against their retrieved records. The entry previously repeated the Zhao study's printed mean differences for the Kawasaki and Tanaka methods as "40.81 g/d and 18.28 g/d", which is the paper's own text but cannot be a sodium excretion figure: the same table reports a measured mean 24-hour excretion of about 3 g/d. The source record already recorded that unit caveat; the entry now does too, and says explicitly that only the direction and gross size of the discrepancy are relied on. The same pass stopped reproducing the study's two non-significant mean biases with confidence intervals that do not contain their own point estimates, and states instead that the measurement point being used is the poor individual-level agreement (intraclass correlation 0.31). Bibliographic identities were re-confirmed from Europe PMC in this pass: Zhao 2021 is PMID 32345383, PMCID PMC11574822, DOI 10.1017/S1368980019005019, Public Health Nutrition 2021;24(4):604-610, and Wu 2022 is PMID 36235755, PMCID PMC9573759, DOI 10.3390/nu14194105, Nutrients 2022;14(19):4105. Neither study measures the shape of the sodium-mortality relationship, and the entry continues to say so.
2026-09-29 additions: Updated the existing potassium-enriched salt strategy with WHO 2025 guidance and explicit population exclusions; no duplicate habit added. Independent approval of this revision remains pending. Publication release 2026-10-04: site owner authorized a bounded AI audit and release of this exact revision. See docs/release-audit-2026-10-04.md for source-check findings and limits. No independent health-professional or peer review is claimed.
Category migration 2026-10-04: removed the separate goals field; categories: Nutrition, Long-term health → Nutrition, Long-term health. Research assessments, sources, guidance, precautions, and evidence review dates are unchanged. This revision awaits publication approval; no new evidence review is claimed.
Publication approval 2026-10-04: site owner explicitly accepted all current category-migration revisions for deployment and requested that remaining limitations be retained in docs/content-review-backlog.md. Approval applies to this exact content revision. No new evidence review, independent expert review, or clinical validation is claimed; existing precautions and disclosures remain in effect.
